[m-dev.] diff: add another regression test

[Fergus Henderson]

On 06-Oct-1998, Thomas Charles CONWAY <conway at cs.mu.OZ.AU> wrote:
> Fergus Henderson, you write:
> > Estimated hours taken: 0.5
> > 
> > tests/hard_coded/Mmakefile:
> > tests/hard_coded/var_not_found.m:
> > tests/hard_coded/var_not_found.exp:
> > 	Add a regression test for a mode analysis bug that resulted
> > 	in a var_not_found error during code generation.
> > 	I don't if the bug was ever fixed, it just happened to
> > 	go away for some unknown reason.
> 
> IIRC, we diagnosed this bug as being some problem to do with complicated
> unification predicates.

After a bit more investigation: the bug was not fixed, it just happened
to no longer cause errors during code generation.  The bug is that
mode analysis generates an incorrect instmap delta, and yes it does
have something to do with complicated unifications: the problem
in this test case occurs at a point where a complicated unification appears
in an implied mode.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>  |  of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3        |     -- the last words of T. S. Garp.